## Super Bowl Game Probability

The Super Bowl game probability is available now at the nytimes.com Fifth Down. Find out what happens to the probability when the each team's passing stats are limited to just those of their starting QBs. Plus, is the Packers' image benefiting from the recency effect?

### 8 Responses to “Super Bowl Game Probability”

1. James says:

.5007? I'm going to go out on a limb and say that's the closest prediction you've ever made.

2. Brian Burke says:

Maybe. I usually wouldn't bother going beyond 50/50, but hey, it's the SB. Plus, the hyper-precision is a juicy hanging curveball for Gregg Easterbrook to knock out of the park.

I mostly just wanted to reassure readers that I was diligent in the calculations.

3. Anonymous says:

How long until someone criticizes Brian in the NYT article comments for refusing to man up and make a real pick?

4. Anonymous says:

Any estimate of the probability this game is decided in overtime ?

5. Ian Simcox says:

The money certainly seems to be following the Packers, they're 57/43 on the latest odds.

I thought this might be a 50:50 prediction based on the efficiency ratings though. I just can't see one major single reason to favour one side over the other - though if I was pushed I'd actually go for Pittsburgh because their rush defense has the best chance to stop the Packers on the ground (this last line is actually related to a research piece I've done but need to send in yet that shows rushing is important to winning up to 3 YPC - above that you don't gain much, but below you lose a lot).

6. Anonymous says:

Brian, when you only look at the Steelers passing efficiency with Ben in the team, do you regress to the mean more severely due to the smaller sample size?

7. Ed Feng says:

It's interesting to look at common opponents of Pittsburgh and Green Bay. They both played Atlanta and the 4 teams in the AFC East. We have a table with results and stats from these games.

http://thepowerrank.com/nfl.html

In short, Green Bay outperformed Pittsburgh in point differential against all but one opponent.

Also, our algorithm has Green Bay as a 2.6 point favorite.

8. Brian Burke says:

Very cool. Thanks for the info, Ed.

Anon-That's a good point about regressing more for BR due to having played fewer games. I didn't think it would matter much because his performance this season is in line with his career mean. Plus, QB passing YPA statistics converge on a relatively stable number after about 10 games played. Regression wouldn't affect it much. The exception is interception rate, which is not as stable. Regressing BR a little further hurts him because his int rate is so low this season. It would tilt things ever so slightly in GB's favor.

### Leave a Reply

Note: Only a member of this blog may post a comment.